JPH10240650A - Document recognition device and recognition processing mehtod - Google Patents

Abstract

PROBLEM TO BE SOLVED: To execute a processing for recognizing picture information on a character, a graphic or the like on an appropriate medium such as paper with a simple operation by providing an information processing unit storing a picture, an information processing unit reading the stored picture to convert it into code information and a network. SOLUTION: Character/picture information inputted from a scanner 2 provided for a document generation work station 1 is transferred to a mass storage device provided in a file station 3 through the network 5 to which the document generation work station 1 is connected. In a recognition work station 4, character/picture information which is thus transferred is recognizer. A recognized result is transferred/stored to/in a file work station 3. The stored recognized result is taken out by the document generation work station 1 one by one as needed. Thus, even a person who is not skilled in a recognition processing operation can easily execute the operation.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a document recognition apparatus using a network configuration for recognizing image information such as characters and graphics printed on a paper document, and a recognition processing method therefor.

[0002]

2. Description of the Related Art Research on various document recognition devices for recognizing characters and figures printed on appropriate paper and inputting information corresponding to the characters and graphics to a document editing device or the like comprising a word processor or the like. Proposals have been made in the past. As such a conventional device, the one disclosed below can be cited as an example. That is, JP-A-63-1
No. 55385, Japanese Patent Application Laid-Open No. 1-18395, Japanese Patent Application Laid-Open No. 1-180081, Japanese Patent Application No. 3-81374.
No., Japanese Patent Application No. 2-333170, Japanese Patent Application No. 3-98014
And the like. The document recognizing devices disclosed therein are a scanner for inputting a document as image information, a character recognizing device for recognizing a corresponding character from the image information, and perform a required display / edit according to a result of the recognition. Display / editing device and the like are integrally configured. Also, "Image Processing Handbook: p. 482" [Editor: Image Processing Handbook Editing Committee (Editorial Chair: Morio Onoe); Publisher: Shokodo Co., Ltd .; The character recognition device disclosed on March 28 (first edition, second printing)] has a configuration in which a required communication control unit is additionally provided. From the viewpoint of the character recognition device, the above-described various It has basically the same configuration as the conventional example.

[0003]

However, since this type of conventional device has the above-described configuration, there are the following problems, for example. (1): Since the document information input from the scanner is taken in as the corresponding image information, the storage capacity of the storage device for storing such image information increases. That is, in order to input a large amount of document information in one conventional device, a storage device having a large storage capacity must be prepared. (2): A scanner must be provided for each workstation, and the conventional apparatus including the scanner and the workstation becomes expensive. (3): Generally, the amount of processing required for character recognition is extremely enormous. For this reason, when character recognition processing is realized by software, this character recognition processing is performed. When the process is being performed, other processes must be interrupted or their priorities must be lowered, and the work efficiency of the operator in such a conventional apparatus is significantly reduced. (4): Even if the character recognition processing is realized by hardware, the device itself for that purpose is expensive, so that the conventional device incorporating such a device becomes extremely expensive. (5) Also, when using the above-described recognition apparatus, the operator of the conventional apparatus needs to acquire new knowledge for recognition processing, and it is extremely difficult for a so-called beginner to use the apparatus. .

The present invention has been made to solve the various problems described above. That is, in a document recognition device configured by connecting a plurality of workstations in a network, a document image is stored in a file station provided with a large-capacity storage function included therein, Therefore, it is not necessary to provide a large-capacity storage device for each workstation. If a scanner is provided in one of the plurality of workstations, and this workstation is dedicated to input (that is, one that is actually operated by an operator), an expensive scanner can be used. It is not necessary to provide for each workstation, and it is possible to suppress the cost of each workstation. Further, one of the plurality of workstations is provided with a character recognition processing function,
Therefore, the load on the workstation actually operated by the operator is correspondingly reduced, and it is not necessary to provide a processing device for character recognition in each workstation. Further, even when the operator is actually operating, the consciousness of using the processing device for character recognition is eliminated, and even a novice user can easily operate.

[0005]

A recognition system according to the present invention is connected to image input means for inputting an image, and includes at least one first storage means for storing an image read by the image input means. Second information processing apparatus (workstation) comprising: an information processing apparatus (workstation); and a conversion unit that reads an image stored in the storage unit and converts the image into code information.
And a network interconnecting the first information processing device and the second information processing device. Further, the document recognition apparatus according to the present invention connects a document creation workstation to which a scanner is connected, a file station for storing image information, and a recognition workstation for recognizing a document image by a network. Comprising: at a document creation workstation,
An image input unit for reading document image information by using a scanner; an image transfer unit for storing document image information as a file in a file station in a document creation workstation; and a file in a recognition workstation. An image extraction unit for extracting document image information from a station; a recognition / conversion unit for recognizing document image information and converting it into corresponding code information in a recognition workstation; and the recognition in a recognition workstation. A recognition result transfer unit for transferring the code information based on the result of the conversion to a file station. Further, a document recognition processing method for a document recognition apparatus according to the present invention includes the steps of:
Confirming the presence or absence of a document image file that is currently being processed, and executing the recognition process depending on the result of this confirmation; the name of the recognized document image file and the date and time of its creation;
Comparing the name of a document image file as a recognition target candidate and the date and time of its creation, and selecting a document image file as a recognition target candidate having a difference between them as an actual recognition target document image file; Is transferred to an appropriate location of the file station, and then a required recognition process is executed. Another document recognition device according to the present invention is configured by further adding an electronic mail transmission workstation.

[0006]

According to the invention, the character image information input from the scanner provided in the document creation workstation is stored in the file station via the network to which the document creation workstation is connected. The data is transcribed to the provided large-capacity storage device. The recognition workstation recognizes the character image information transcribed in this way, and the result of the recognition is transferred and stored to the file station.
The recognition results stored in this way are individually retrieved by the document creation workstation as needed. This eliminates the need for a scanner, a large-capacity storage device, and recognition software (or hardware) at each document creation workstation. Thus, the overall cost of the apparatus is reduced, and there is no need to provide special hardware and software at the document creation workstation directly used by the operator. However, the operation can be easily performed. Further, according to the present invention, since the recognition result for the document information as the recognition target is transmitted to the operator as an electronic mail, the operator waits for the transmission of the recognition result. While you can do other work. Further, according to the present invention, since the designation of the area desired to be recognized can be executed by the operator of the document creation workstation, for example, a document having a mixture of tables and figures or a document having columns is used. Can be treated as a recognition target.

[0007]

FIG. 1 is a schematic configuration diagram showing a network-structured document recognition apparatus according to a first embodiment of the present invention. The network-structured document recognition apparatus of the first embodiment shown in FIG. 1 includes a document creation workstation 1 directly operated by an operator, and is exclusively connected to the document creation workstation 1. Scanner 2, large-capacity storage device (not shown)
And a recognition workstation 4 for recognizing document image information stored in a storage device in the file station 3. These means are mutually connected via a network 5. It is connected. FIG. 2 is a schematic functional configuration diagram for functionally explaining the recognition processing in the first embodiment. FIG. 3 is a functional explanatory diagram of the storage device in the file station 3 in the first embodiment. FIG. 4 is a view showing an example of a recognized registration file in the first embodiment. FIG. 5 is a flowchart showing the mode of the recognition processing in the first embodiment. Hereinafter, a first embodiment of the present invention will be described with reference to FIGS.

[0008] The operator is a workstation 1 for creating a document.
Is processed by the image input unit 11 and the dedicated scanner 2 is operated to digitally input required document information described on a suitable medium such as paper as corresponding image information. At this time, the original image information corresponding to the previous document information is temporarily stored in a predetermined storage device (not shown) in the document creation workstation 1, but the following operation can be performed. That is, a target image is displayed on an appropriate display device 141 by the control / operation unit 14 in the document creation workstation 1, and whether or not this image is a regular input image has an appropriate image quality. It is also possible to make the operator confirm whether or not there is, and then re-enter the confirmation.

In an image transfer section 12 in the document creation workstation 1, a document image as a recognition target input from the scanner 2 is transmitted via a network 5.
The data is transferred to the image storage unit 31 in the file station 3. As shown in FIG. 3, a storage device having a required storage capacity in the file station 3 stores the image storage unit 31 for storing a document image as a recognition target, and stores the recognition result. Result storage unit 3 for performing
And is appropriately used depending on the purpose. Here, a specific operation performed by the operator of the document creation workstation 1 is, for example, as follows. Assuming that information on a document image to be recognized is displayed on the display screen of the display device 141 as a corresponding appropriate icon, the operator can use a suitable input device 142 such as a keyboard or a mouse. The selection is made using an appropriate pointing device 143 as described above. Then, the icon of the selected document image is displayed on the display device 141 as a corresponding icon.
Is transcribed to the image storage unit 31 displayed on the display screen. In this way, a document image as a desired recognition target can be transferred to the image storage unit 31 in the storage device in the file station 3. Such an operation instruction from the operator is transmitted to the image transfer unit 12 via the control / operation unit 14.

Next, the document image to be recognized, which is transcribed and stored in the image storage unit 31 in the storage device in the file station 3, is recognized by the recognition workstation 4. FIG. 5 shows a form of the recognition processing here as a flowchart. In addition,
The recognition process shown here is a process that is started at a certain time, for example, the process is started at 0 minutes every hour. However, it is needless to say that the start time and the number of operations of the process are It can be arbitrarily selected and set. Now, if an instruction to start a predetermined recognition process is issued, first,
It is confirmed whether or not there is a recognition processing file as a file undergoing the recognition processing (step 420). The reason why this step 420 is necessary is as follows. That is, as described above, the processing illustrated here is started at a certain time (for example, started at 0 minute every hour).
When the recognition processing for a certain object takes a long time (for example, a time longer than one hour) because the processing is regarded as the processing, the next recognition processing is performed prior to the end of the recognition processing. Processing may start. In such a case, the recognition processing for the same object is performed in parallel and doubly, and extra processing is performed accordingly. For such a reason, a confirmation step of step 420 as described above is provided, and when the existence of the file being recognized is confirmed (when the confirmation result is Y), the recognition processing is performed. The end is instructed, and the start of double recognition processing for a document image for which recognition processing has already been started is prevented. When it is confirmed in step 420 that the file under recognition processing does not exist (when the result of the confirmation is N), a file under processing is created in step 421. Subsequently, by the operation of the image extracting unit 41 in the recognition workstation 4 (FIG. 2), all the images to be recognized are recognized from the image storage unit 31 in the storage device in the file station 3 via the network 5. The information corresponding to the document image is taken out and taken into a storage device (not shown) in the recognition workstation 4 (step 422).

Here, regarding the transferred document image,
The documents are sequentially subjected to the recognition processing one by one. First, it is confirmed whether or not there is a document image to be transferred (step 423). When the existence of the transferred document image is confirmed (when the result of the confirmation is Y), in the following step 424, it is determined whether or not the corresponding transferred document image has already been recognized. Is confirmed. Here, whether or not the recognition has been completed is performed as follows. That is, as in step 427 of registration, which will be described later, when the recognition process for a given target is completed, the corresponding recognized registration file (which is
4, the file name of the transferred document image and the creation date and time (the creation date and time is the date and time read by the scanner) are registered. Referring now to FIG.
(NYamamoto, NShimizu) is registered as the file name of the document image, and "Dec.
26 19:50 (Jan 10 19:50, Jan
11 09:36) "corresponds to the creation date and time. Therefore, when the recognized file name registered in the recognized registration file and the creation date and time are the same before the last time, this is already recognized (when the confirmation result is Y, ), The process returns to the previous step 423. Here, the registered recognized file names and their creation dates and times are compared for the following reasons. That is, for example, when there are two document image files having different contents, they may be given exactly the same file name, but these are compared only with the registered recognized file names. At times, it is excluded from the target of the recognition processing. In order to prevent such a situation from happening, in addition to the registered recognized file name, the created date and time are also compared, so that it is necessary to confirm the identity of the two. Have been. Now, as a result of the confirmation in step 424, when it is determined that the transferred document image has not been recognized yet, in step 425 following this, the recognition / conversion unit 42 of the recognition workstation 4 A required recognition process is performed by the operation, and predetermined code information corresponding to the result of the recognition is converted and generated. In the recognition step 425, a character area in the recognition target is extracted, and character recognition is performed there. There is no specific method of character recognition here,
For example, the method disclosed in Japanese Patent Application No. 3-195100 can be used. Subsequently, the predetermined code information, which is the result of the above-described character recognition, is transferred to the file station 3 in the storage device via the network 5 by the operation of the recognition result transfer unit 43 in the recognition workstation 4. The data is transferred to the result storage unit 32 (step 426). Subsequent step 42
At 7, the file name of the document image information file to be recognized and its creation date and time are registered in the recognized registration file, as described above. Then, the document image information registered as a recognition target as described above is deleted (step 428).

As long as there is a transfer image to be recognized, the same processing as described above is repeated. Then, when the processing for all the transfer images is completed (when the result of the processing in step 423 is N), the file being processed is deleted (step 429) so that new recognition processing can be dealt with. Is done.

The image to be recognized is a file station 3
When an appropriate time has elapsed since the data was transcribed to the storage device, the operator checks the recognition result storage unit 32 in the storage device in the file station 3 so that the desired recognition result is stored. Check if it is. When the storage of the desired recognition result is confirmed, the desired recognition result is extracted and used for document creation by the operation of the recognition result extracting unit 13 in the document creation workstation 1.

Although the first embodiment of the present invention has been described in detail, various changes and modifications as described below can be made. (1) In the first embodiment, one document creation workstation with a scanner, one file station with a large-capacity storage device, and one recognition workstation are connected to the network. Although it is configured, it is not limited to such a configuration, and a plurality of components can be connected and configured if necessary. Further, a single workstation having a plurality of functions can be provided. (2) In the first embodiment, both reading of image information and taking in of the recognition result are performed by only one document creation workstation. Can be done at the workstation. For example, one workstation dedicated to inputting image information may be connected to the network, and this dedicated workstation may be used by all operators involved, and the recognition result may be used at each operator's own workstation. . (3) In the first embodiment, the storage device in the file station 3 is divided into the image storage unit 31 and the recognition result storage unit 32. However, the present invention is not limited to this. , And the image and the result of its recognition can be stored in the same storage unit. In this case, the image for which the recognition processing has been completed can be deleted. (4) In the first embodiment, the information about the document image to be recognized is directly input from the scanner. However, the present invention is not limited to this. For example, the information is temporarily stored in a recording medium such as an optical disk file. It is also possible to search and recognize a document image to be reused from recorded ones. (5) Further, in the first embodiment, the recognition of characters is described as an example of the recognition processing. However, the present invention is not limited to this. Raster to convert to information /
A vector conversion function, a graphic recognition function for recognizing a basic figure (for example, a triangle or a quadrangle), or a document recognition function for a document in which characters / graphics are mixed may be provided.

FIG. 6 is a schematic block diagram showing a network-structured document recognition apparatus according to a second embodiment of the present invention. The network-structured document recognition apparatus in the second embodiment shown in FIG. 6 includes a document creation workstation 1 directly operated by an operator, and is exclusively connected to the document creation workstation 1. Scanner 2, a file station 3 provided with a large-capacity storage device (not shown), and a recognition work for recognizing document image information stored in a storage device in the file station 3. A station 4 and a workstation 6 for transmitting e-mail required electronic information to a designated destination in response to a predetermined command, these means being connected to each other via a network 5 Have been. FIG. 7 is a schematic functional configuration diagram for functionally explaining the recognition processing in the second embodiment. FIG. 8 is a flowchart for illustrating an operation mode of the electronic mail transmitting workstation 6 in the second embodiment. The storage device in the file station 3 in the second embodiment is similar in function to that of the first embodiment described above. Please refer to FIG. 3 as an explanatory diagram for the embodiment. The recognized registration file in the second embodiment is also the same as that of the first embodiment described above. Therefore, if necessary, FIG. 4 as an illustration for the first embodiment is used. See Further, the mode of the recognition processing in the second embodiment is also the same as that of the first embodiment described above. Therefore, FIG. 5 as a flowchart for the first embodiment may be used as needed. refer. Hereinafter, FIGS. 6 to 8
3 to 5 relating to the first embodiment described above.
A second embodiment of the present invention will be described with reference to FIG.

The operator operates his or her own document creation workstation 1, processes the image input unit 11 thereof, activates the dedicated scanner 2, and operates a dedicated scanner 2 so that the required information written on a suitable medium such as paper is written. The document information is digitally input as the corresponding image information. At this time, the original image information corresponding to the previous document information is stored in the document creation workstation 1.
Is temporarily stored in a predetermined storage device (not shown), but the following operation can also be performed. That is, a target image is displayed on an appropriate display device 141 by the control / operation unit 14 in the document creation workstation 1, and whether or not this image is a regular input image has an appropriate image quality. It is also possible to make the operator confirm whether or not there is, and then re-enter the confirmation.

In the image transfer unit 12 in the document creation workstation 1, a document image to be recognized input from the scanner 2 is transmitted via the network 5.
The data is transferred to the image storage unit 31 in the file station 3. As shown in FIG. 3, a storage device having a required storage capacity in the file station 3 stores the image storage unit 31 for storing a document image as a recognition target, and stores the recognition result. Result storage unit 3 for performing
And is appropriately used depending on the purpose. Note that the recognition result accumulation unit 32 functions to accumulate the recognition result when the transfer process has failed. Here, specific operations performed by the operator of the document creation workstation 1 are, for example, as follows. Assuming that information on a document image to be recognized is displayed on the display screen of the display device 141 as a corresponding appropriate icon, the operator can use the input device 142 such as a keyboard or the mouse such as a mouse. The selection is made using an appropriate pointing device 143. Then, the icon of the selected document image is displayed on the display device 1 as a corresponding icon.
The data is transcribed to the image storage unit 31 displayed on the display screen 41. In this way, a document image as a desired recognition target can be transferred to the image storage unit 31 in the storage device in the file station 3.
Such an operation instruction from the operator is transmitted to the image transfer unit 12 via the control / operation unit 14.

Next, the document image to be recognized, which is transcribed and stored in the image storage unit 31 in the storage device in the file station 3, is recognized by the recognition workstation 4. FIG. 5 shows a form of the recognition processing here as a flowchart. In addition,
The process shown here is a process that is started at a certain time, for example, the process is started every hour at 0 minutes. Of course. If an instruction to start a predetermined recognition process is issued, first, it is confirmed whether or not there is a file under the recognition process as a file under the recognition process (step 420). The reason why this step 420 is necessary is as follows. That is, as described above,
Since the process illustrated here is assumed to be started at a certain time (for example, started at 0 minute every hour), it takes a lot of time for the recognition process for one object. If it takes longer than one hour, for example, the next recognition process may start before the recognition process ends. In such a case, the recognition processing for the same object is performed in parallel and doubly, and extra processing is performed accordingly. For such a reason, a confirmation step such as the above-described step 420 is provided, and when the presence of the file being recognized is confirmed (when the confirmation result is Y), the recognition processing is performed. The end is instructed, and the start of double recognition processing for a document image for which recognition processing has already been started is prevented. When it is confirmed in step 420 that the file under recognition processing does not exist (when the result of the confirmation is N), a file under processing is created in step 421. Subsequently, by the operation of the image extracting unit 41 in the recognition workstation 4, the image storage unit 31 in the storage device in the file station 3 supports all document images as recognition targets via the network 5. The information to be retrieved is taken out and stored in a storage device (not shown) in the recognition workstation 4 (step 422).

Here, regarding the transferred document image,
The documents are sequentially subjected to the recognition processing one by one. First, it is confirmed whether or not there is a document image to be transferred (step 423). When the existence of the transferred document image is confirmed (when the result of the confirmation is Y), in the following step 424, it is determined whether or not the corresponding transferred document image has already been recognized. Is confirmed. Here, whether or not the recognition has been completed is performed as follows. That is, as in step 427 of registration, which will be described later, when the recognition process for a given target is completed, the corresponding recognized registration file (this is illustrated in FIG. 4) is The file name of the transferred document image and the creation date and time (the creation date and time is the date and time read by the scanner) are registered. Therefore, when the recognized file name registered in the recognized registration file and the creation date and time are the same before the last time, this is already recognized (when the confirmation result is Y, ), The process returns to the previous step 423. Here, the registered recognized file name and its creation date and time are compared with each other for the reason already described in the first embodiment, but will not be described again here. If it is determined in step 424 that the transferred document image has not been recognized yet, the process proceeds to step 42.
At 5, the recognition / conversion unit 42 in the recognition workstation 4 performs a required recognition process, and generates predetermined code information corresponding to the recognition result. In the recognition step 425, a character area in the recognition target is extracted, and character recognition is performed there. Although there is no specific method for character recognition here, for example, the method disclosed in Japanese Patent Application No. 3-195100 can be used. Subsequently, predetermined code information as a result of the character recognition described above is transmitted to the network by the following operation of the recognition result transfer unit 43A associated with both the recognition workstation 4 and the electronic mail transmission workstation 6. Via 5
The document is transferred to the document creation workstation 1 corresponding to the predetermined operator (step 426). Here, FIG.
The operation of the recognition result transfer will be described with reference to the flowchart of FIG. First, when the recognition result transfer process is started by any operation trigger, an operator (user) name corresponding to a file name to be recognized is extracted (step 81). In the next step 82, the operator (user) name is confirmed as the address. Following this, it is determined whether there is an address (step 83). If the result of this determination is affirmative (Y), an electronic mail is transmitted to the corresponding destination (step 84). If the result of the above determination is negative (N), for example, the file The electronic mail is transferred to the recognition result accumulation unit 32 in the use station 3 (step 85). After such processing is completed, the transfer processing of the recognition result ends as a whole. Such a transfer process of the recognition result can also be described as follows. That is, the recognition result transfer unit 43A extracts the user name as the operator corresponding to the file name as the recognition target, and uses this user name as the address to send a command to transmit the recognition result file to the electronic mail transmission work. Transfer to station 6. Workstation 6 for electronic mail transmission
When receiving the above-mentioned electronic mail transmission command from the recognition workstation 4, the specified mail transmission command is sent to a specific one of the document creation workstations 1 connected to the network 5. Transfer files. At this time, if the transmission of the required e-mail fails due to troubles such as an incorrect user name or an unknown destination (user or workstation to be transferred), the transfer destination of the above-described recognition result file is ,
The result is stored in the recognition result storage unit 32 in the storage device in the file station 3. Next, step 427 (of FIG. 5)
In, as described above, the file name of the document image information file to be recognized and its creation date and time are
Registered in the recognized registration file. Then, the document image information registered as a recognition target as described above is deleted (step 428).

As long as there is a transfer image to be recognized, the same processing as described above is repeated. Then, when the processing for all the transfer images is completed (when the result of the processing in step 423 is N), the file being processed is deleted (step 429) so that new recognition processing can be dealt with. Is done.

The image to be recognized is a file station 3
Is transferred to the storage device, the operator waits for the recognition result regarding the recognition target image to be transmitted as an electronic mail while engaging in other tasks. Assuming that a desired electronic mail has been transmitted, a change occurs on the display screen of the display unit of the document creation workstation 1 used by the operator, indicating that the corresponding recognition result has been transmitted. Examples of the change on the display screen include a blink of an icon corresponding to the electronic mail and a change in its shape.
In this way, the fact that the desired recognition result has been transmitted is clearly displayed through the visual sense, so that the operator does not need to worry about whether or not the predetermined recognition processing has been completed, and The time until the recognition result is transmitted can be fully utilized for other tasks. Then, when the transmission of the desired recognition result is confirmed, the desired recognition result is extracted from the icon representing the electronic mail by the operation of the recognition result extracting unit 13 in the workstation 1 for document creation, and is used for document creation. To be used.

Although the second embodiment of the present invention has been described in detail, the following changes or modifications can be made in addition to the case of the above-described first embodiment. (6): In the second embodiment, the recognition result for the document information to be recognized is transmitted to the operator as an electronic mail. However, the present invention is not limited to this, and the file transfer function provided on the network is not limited to this. And can be directly transferred to a predetermined storage area in a specific workstation.

FIG. 9 is a schematic configuration diagram showing a network-structured document recognition apparatus according to a third embodiment of the present invention. The network-structured document recognition apparatus of the third embodiment shown in FIG. 9 includes a document creation workstation 1 including an image input unit directly operated by an operator, and a document creation workstation 1. A scanner 2, which is exclusively connected to a scanner 2, and a recognition workstation 4 including a recognition / conversion unit for recognizing (and converting to corresponding code information) the document image information from the document creation workstation 1, These means are connected to each other via a network 5. The device according to the third embodiment is the same as the device according to the first embodiment described above, except for the following improvements. That is, an improvement has been made so that a recognition target area for a document image in which characters and graphics are mixed can be easily specified without providing a special function or device for document image recognition on the document creation workstation side. It was done. FIG. 10 is a schematic functional configuration diagram for functionally explaining the recognition processing in the third embodiment.
Hereinafter, a third embodiment of the present invention will be described with reference to FIGS. 9 and 10 as appropriate.

The operator is a workstation 1 for creating a document.
Is operated, and by operating the scanner 2,
The image input unit 11A performs a process to digitally input required document information including characters and graphics described on a suitable medium such as paper as corresponding image information. The image information input in this manner is displayed on a suitable display device 141 as a target image by the control / operation unit 14B in the document creation workstation 1.

The image editing unit 12A in the document creation workstation 1 has an input device 14 such as a keyboard.
By interactively editing an image using a pointing device 143 such as a mouse 2 or a mouse, a recognition area as a target is specified. Although there are various methods for specifying the recognition area here, what is important is that, at the stage of extracting the area, it is possible to distinguish between an original image element and an image element added later. That is. Here, as an example,
A method of surrounding a region desired to be recognized with an appropriate rectangular frame will be described. In general, it is common that an appropriate rectangular frame exists in a given document.However, when this rectangular frame is digitized by a scanner or the like, noise or unwanted inclination occurs. Some inconveniences may occur, such as generation or non-constant thickness. Further, regarding the rectangular frame on the display screen drawn by the image editing unit 12A, it is possible to specify the recognition target area using a rectangular frame having a strict line width such as, for example, one dot width. The drawing function of the rectangular frame performed by the image editing unit 12A can be generally used in the document creation workstation 1, and is not added again for recognizing a document image. Here, FIG.
With reference to FIG. 1, an example of the specification of a rectangular frame in such a document image is illustrated. In FIG. 11,
Although not limited to this, a predetermined document image display surface 110 includes a # 1 rectangular frame 111 and a # 2 rectangular frame 11
2 and # 3 rectangular frame 113 are illustrated as being present. The designation of the rectangular frame here is performed by the operator at the document creation workstation 1 using the display device 14.
Using the pointing device 143 to specify two vertices of a diagonal of a desired rectangular frame (for example, an upper left corner and a lower right corner of the rectangular frame) while viewing the corresponding document image on the display screen of (1). Done by Note that, in order to specify the recognition area here, in addition to actually defining a rectangular frame, the two vertices must be specified using a key symbol (",") or a cross symbol (+). Can also. By the way, the image file in which the recognition target area is designated in this way is sent to the image output unit 13 in the document creation workstation 1.
A transfers the image data to the image input unit 4A of the recognition workstation 4. As described above, the document creation workstation 1 and the recognition workstation 4
Is connected via the network 5, the transfer mode here is direct transfer using an electronic mail function, or as described in the first embodiment, Via a file station (not shown).

The image file received by the image input unit 4A is sent to a recognition area extracting unit 4B at the next stage. The recognition area extraction unit 4B extracts a recognition target area specified by the image editing unit 12A of the document creation workstation 1. Here, referring to FIG. 12, there is exemplified a mode of recognizing a region surrounded by a rectangular frame having a strict line width of 1 dot. 12, a predetermined document image display surface 110 is illustrated as in the case of FIG. 11 described above, but only the # 1 rectangular frame 111 therein is taken up. On such a document image display surface 110, scanning is sequentially performed from the upper left corner in a so-called TV system. Then, as schematically indicated by the arrow 111A, the # 1 rectangular frame 11
When the first black pixel is found, the coordinate value at this point is first stored. Following this, black pixels are tracked rightward in the figure (ie, rightward along the top side of the # 1 rectangular frame 111). From the point where the next pixel is a white pixel (that is, the bending point of the arrow 111B), the black pixel is tracked downward (downward along the right side of the # 1 rectangular frame 111).
At this time, the number of black pixels for each tracking direction (here, the number of black pixels in the right direction and the number of black pixels in the downward direction) is stored, and the point where the next pixel is a white pixel (arrow 111).
C (the bending point of C) to the left of the figure (# 1 rectangular frame 111).
(To the left along the lower side of). When tracking in the left direction, tracking can be performed by the number of black pixels known in the previous tracking in the right direction. And from the bending point of the last arrow 111D,
The black pixel is tracked upward (upward along the left side of the # 1 rectangular frame 111). At this time, it is possible to track only the number of black pixels known in the downward tracking. When a known number of black pixels are tracked along the lower side and left side of the # 1 rectangular frame 111, and a white pixel is found, the currently tracked rectangular frame is a regular one. Is not considered. Further, as described above, since the line width of the rectangular frame is strictly defined as one dot, when the black pixel is tracked, it is in contact with the left and right sides in the tracking direction. If there is a black pixel of one dot or more, this is also regarded as a non-regular rectangle, the tracking is stopped, and the process proceeds to the extraction of the next rectangular frame.

In such tracking of black pixels, the rectangular frame sometimes crosses the noise component portion of the original image, so to speak, and may not have a strict one-dot line width. Therefore, considering that such a case may occur, even if there are black pixels on the left and right sides in the traveling direction of tracking the rectangular frame, such a state is within a certain length range. If there is a certain thing, its existence may be allowed. Also, since the rectangular area surrounded by the rectangular frame cannot be extremely small, when the length of the upper side is smaller than a certain value in the first tracking of the rectangular frame, the normal The following tracking can be stopped by assuming that the frame is not a rectangular frame. Note that such a method can be similarly applied to a rectangular frame having a line width of 2 dots or more.

When a plurality of rectangular areas are extracted in one document image display surface, the rectangular areas are sent to the recognition / conversion unit 4C in the recognition workstation 4 in any order. Transferring can be a problem. Referring to FIG. 13, an example in which three rectangular regions are included in a certain document image display surface 110 is illustrated. In FIG. 13, # 1 surrounded by a rectangular frame #
One region 111R, # 2 region 112 surrounded by a # 2 rectangular frame
R and a # 3 region 113R surrounded by a # 3 rectangular frame are shown as being present. In FIG. 13, the coordinates (0, 0, 1) of the left shoulder of the document image display surface 110 are shown.
The point 0) is the starting point of tracking, the arrow x is the horizontal tracking direction in the figure, and the arrow y is the vertical tracking direction in the figure. In FIG. 13, when a plurality of rectangular areas are sequentially extracted, the upper side of the # 3 area 113R is closer to the upper side of the # 2 area 112R when viewed from the lower side of the # 1 area 111R. # 1 area 111R, # 3 area 11
3R and # 2 area 112R are extracted in this order.
However, from the viewpoint of a human operator who accepts these recognition results, the # 1 area 111R and the # 2 area 1
It is preferable that the recognition result is returned in the order of 12R and # 3 area 113R. Therefore, as the post-processing of the rectangular area extraction described above, the # 2 area 112R and the # 3 area 113
In the case of a rectangular area such as R, which is arranged side by side in the figure (that is, the y-coordinates of both overlap),
The rectangular area on the left side (that is, the x coordinate is smaller) is transferred to the recognition / conversion unit 4C first. The recognition / conversion unit 4C extracts a required character area from the rectangular area obtained by the recognition area extraction unit 4B, performs character recognition, and converts the character area into a character code corresponding to the recognition result. Although the method of character recognition here is not particularly limited, for example, Japanese Patent Application No. Hei.
No. 5100 can be applied. And
The character code as a result of the character recognition described above is sent to the document creation workstation 1 via the network 5.
Is transferred to the recognition result input unit 14A. Then, the operator of the document creation workstation 1 uses the above recognition result to create a document via the control / operation unit 14B.

Although the third embodiment of the present invention has been described in detail, the present invention is not limited to this embodiment, and arbitrary and desired changes and modifications can be made without departing from the spirit of the present invention. That is, the third
In the embodiment, the document creation workstation 1
Although the example in which only the recognition workstation 4 is connected to the network 5 is shown in FIG.
It is possible to use a file station as in the second embodiment, or to use an electronic mail transmission workstation as in the second embodiment.

[0030]

As described above in detail, according to the network configuration recognition system (for example, a document recognition device) according to the present invention, document image information input from a scanner by an operator is stored in a common file. Operating in a large-capacity storage device provided in the storage station, a large-capacity storage device is individually provided for a document creation workstation prepared for each operator. There is no need to provide In addition, at least one scanner for inputting document image information only needs to be provided in the network-structured document recognition apparatus of the present invention, and is provided separately for all document creation workstations for each operator. No need. Further, since the recognition of the document image in the present invention is performed by the recognition workstation in the document recognition device having the network configuration, the load for the document creation workstation for each operator is not applied, and the operator's work is not required. The work is not interrupted by the recognition operation.
Further, there is no need to connect recognition hardware or install software to the document creation workstation. That is, it is not necessary to additionally provide the above-mentioned special hardware and software for the document creation workstation for each operator, and the document image recognition function can be provided by existing means or devices. Can be fully utilized. As for the operator himself, since all the means or devices for direct operation are already existing, there is no need to learn a particularly new operation method, and even a beginner can perform document image recognition. Functions can be used easily and reliably. Furthermore, according to the present invention, since the operator of the document creation workstation can specify an area desired to be recognized, for example, a document having a mixture of tables and figures or a document having columns is used. Can be treated as a recognition target, and by designating only a part necessary for the recognition operation, it is possible to avoid useless use of computer resources. In such a case, it is only necessary to divert general-purpose functions without specially providing recognition hardware or software on the document creation workstation. There is no need to add special functions of.

[Brief description of the drawings]

FIG. 1 is a schematic configuration diagram illustrating a document recognition device having a network configuration according to a first embodiment of the present invention.

FIG. 2 is a schematic functional configuration diagram for functionally explaining a recognition process in the first embodiment.

FIG. 3 is a functional explanatory diagram of the storage device according to the first embodiment.

FIG. 4 is an exemplary diagram of a recognized registration file in the first embodiment.

FIG. 5 is a flowchart illustrating an aspect of a recognition process in the first embodiment.

FIG. 6 is a schematic configuration diagram illustrating a network-structured document recognition device according to a second embodiment of the present invention.

FIG. 7 is a schematic functional configuration diagram for functionally explaining a recognition process in the second embodiment.

FIG. 8 is a flowchart for illustrating an operation mode of the electronic mail transmitting workstation 6 in the second embodiment.

FIG. 9 is a schematic configuration diagram illustrating a document recognition device having a network configuration according to a third embodiment of the present invention.

FIG. 10 is a schematic functional configuration diagram for functionally explaining a recognition process in the third embodiment.

FIG. 11 is a view showing an example of a mode of designating a rectangular frame of a document image in the third embodiment.

FIG. 12 is a diagram exemplifying a mode of recognizing a region surrounded by a rectangular frame having a certain line width in the third embodiment.

FIG. 13 is a diagram illustrating an example in which three rectangular areas are included in a certain document image display surface 110 in the third embodiment.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Workstation for document preparation, 11 ... Image input part, 12 ... Image transmission part, 13 ... Recognition result extraction part, 1
4 Control / operation unit 2 Scanner 3 File station 31 Image storage unit 32 Recognition result storage unit 4 Recognition workstation 41 Image retrieval unit 42 Recognition / conversion unit 43 ... Recognition result transfer unit, 5
... Network, 6 ... E-mail sending workstation.

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁶ Identification code FI H04L 12/58

Claims (7)

[Claims] 1. A document image input device for inputting a document image, a recognition device for recognizing a document image input by the document image input device and converting the document image into corresponding code information, Address extracting means for extracting an address corresponding to the document image obtained, and electronic mail transmitting means for transmitting, as an electronic mail, the code information converted by the recognizing means to the address extracted by the address extracting means. A document recognition device characterized by the following. 2. A document processing information processing device connected to a scanner, a file information processing device for storing a document image, and a recognition information processing device for recognizing a document image are connected via a network. In the document recognition apparatus, the information processing device for document creation includes: an image input unit that inputs a document image by a scanner; and a document image input by the image input unit is transferred to the file information processing device as a file. A document image transfer unit, the file information processing device includes a storage unit that stores the document image transferred by the document image transfer unit in the document creation information processing device, and the recognition information processing device. A document image extracting unit for extracting a document image stored in a storage unit in the file information processing device; Recognizing means for recognizing a document image taken out by the document image taking out means and converting it into corresponding code information; destination taking out means taking out a destination corresponding to the document image taken out by the document image taking out means; For the destination retrieved by means,
A document recognizing apparatus, comprising: an electronic mail transmitting unit that transmits the code information converted by the recognizing unit as an electronic mail. 3. A document processing information processing device connected to a scanner, a file information processing device for storing a document image, and a recognition information processing device for recognizing a document image are connected via a network. In the document recognition apparatus, the document creating information processing device includes: an image input unit that inputs a document image by a scanner; and a document image input by the image input unit is transferred to the file information processing device as a file. A document image transfer unit, the file information processing device includes a storage unit that stores the document image transferred by the document image transfer unit in the document creation information processing device, and the recognition information processing device. A document image extracting unit for extracting a document image stored in a storage unit in the file information processing device; Recognizing means for recognizing a document image taken out by the document image taking out means and converting it into corresponding code information; and operator name taking out means taking out an operator name corresponding to the document image taken out by the document image taking out means. When the operator name extracted by the operator name extracting unit is set as a destination, a determining unit that determines whether there is the destination, and when the determining unit determines that there is a destination,
A document recognizing apparatus comprising: an electronic mail transmitting unit that transmits, as an electronic mail, the code information converted by the recognizing unit to the destination. 4. The information processing device for recognition, when the determination unit determines that there is no destination,
4. The document recognition apparatus according to claim 3, further comprising a recognition result transmitting unit that transfers the code information converted by the recognition unit to the file information processing device. 5. An information processing apparatus for creating a document to which a scanner is connected, an information processing apparatus for storing a document image, and an information processing apparatus for recognizing a document image are connected by a network. In the document recognizing apparatus, the information processing apparatus for creating a document inputs a document image by a scanner, transfers the input document image as a file to the information processing apparatus for a file, Storing the document image for document creation and transferred by the information processing device, the recognition information processing device retrieves the document image stored in the file information processing device, and retrieves the retrieved document image. Recognize and convert to the corresponding code information, take out the destination corresponding to the retrieved document image, and retrieve the retrieved destination And transmitting the converted code information as an electronic mail. 6. A document processing information processing device connected to a scanner, a file information processing device for storing a document image, and a recognition information processing device for recognizing a document image are connected via a network. In the document recognition apparatus, the information processing device for document creation includes: an image input unit that inputs a document image by a scanner; and a document image input by the image input unit is transferred to the file information processing device as a file. A document image transfer unit, the file information processing device includes a storage unit that stores the document image transferred by the document image transfer unit in the document creation information processing device, and the recognition information processing device. A document image extracting unit for extracting a document image stored in a storage unit in the file information processing device; Recognizing means for recognizing a document image taken out by the document image taking out means and converting it into corresponding code information; destination taking out means taking out a destination corresponding to the document image taken out by the document image taking out means; A document recognition apparatus, comprising: an electronic mail transmitting unit that transmits, as an electronic mail, an end of recognition by the recognition unit to a destination extracted by the unit. 7. A document processing information processing apparatus connected to a scanner, a file information processing apparatus for storing a document image, and a recognition information processing apparatus for recognizing a document image are connected via a network. In the document recognizing apparatus, the information processing apparatus for creating a document inputs a document image by a scanner, transfers the input document image as a file to the information processing apparatus for a file, Storing the document image transferred by the document creation information processing device, the recognition information processing device retrieves the document image stored in the file information processing device, and recognizes the retrieved document image And converts it into the corresponding code information, extracts the destination corresponding to the extracted document image, and On the other hand, a document recognition method characterized by transmitting the completion of recognition as an electronic mail.